Forum Moderators: DixonJones
It would collect the following data (using ECMAScript):
* Flash version
* Google Gears version
* Screen resolution an color depth
* UA-related information (UA string, version, OS, platform, etc.)
* Plugins installed (Mozilla-based)
* Referrer URI
* Whether the browser is in Quriks or Standard mode
* Local time and timezone information
* Support of ActiveX and XHR.
* Result of some basic DOM tests to check DOM support.
In addition, the use of server logs would provide with:
* IP address
* country of origin
* methods to check whether the user is a new or a recurring visit.
I would like to know what you think about such a system. I'd be glad if you could tell me what kind of information is actually essential to you, or that you would like to be able to know and your current system doesn't provide.
Thanks in advance
JavaScript (and JScript) are currently aimed to be compatible with ECMAScript although they may provide additional features that are not defined in the ECMAScript specification per se.
Actually, color depth is a highly valuable data. Imagine that you find out that 20% of your visitors only support up to 256 colors.
If you find out your users only support 256 colors you're either running the internet in the early 90s or you're in the Twilight Zone.
The size of the display device is actually more interesting these days than the color depth.
does anyone know how the average user in the developing countries is equipped
No clue but everything would be kind of useless to them if they only had 256 colors.
I would assume smaller monitors such as 1024x768 but high color.
Think about it, with low colors any photography, video or flash content would be intolerable to watch which would make the whole internet experience less that fun.
I can see why you'd want to track all of this if you are in charge of a web site and you want to be able to apply the latest apps and such, but you want to be sure your visitors actually can use them. And if you're selling/making plug-ins or something similar, you'd like to see what plug-ins are already being used, and watch the trends.
And I know for a fact that detecting some of the things you're after will happen only if you get to know those things extremely well, to learn the clues that give them away (some are very simple, others not so much). That's a valuable exercise to go through.
And if you don't want to install or pay for an existing analytics program, or if you feel they will be exploiting you, then writing your own program will have its own kind of payoff.
At the same time I hope you are aware that you are probably re-inventing a wheel that has been invented many times, because a lot of analytics data collection programs collect most of what you're describing, and so do simple server logs.
Also, you are talking about only the data collection piece of analytics. The further parts of analytics (sessionizing, organizing, presentation, and actual analyzing) are where you'll fall short unless you have a very thorough understanding of what all the other tools do and why. Your data will probably end up being "analyzed" by something like Excel or LogParser, and even then won't be able to be useful unless the data is summarized into visits (not hits).
That said, if you can write bits of code for detecting some of these more obscure things, and if your bits of code can be fit into other programs' analytics page tagging so they can add to what's already being collected by them, you'll do some people a service if you decide to share. It'll be difficult to add to what Google Analytics offers because Google Analytics wouldn't be very adaptable about the additional data types (prove me wrong on that and I'll be happy!), but the cost-based, more sophisticated programs like Omniture, Coremetrics, and especially WebTrends (which is extremely adapatable to additional collected information) are those that will allow their users to see the additional information in their reports. I'm not sure about Yahoo Analytics, I haven't started playing with it yet.
I realize that I haven't answered your original question at all, that is, what facts about the visitor's browser and system are important to collect. Sounds like you have a handle on that with your own ideas. If you can collect it, do it, as long as you understand that most of your list is already being done by most existing analytics programs. But somebody somewhere will need to know about Google Gears support and so on!
Isn't the goal of studying web site analytics to improve your site's goals? If so, how is this crap going to help do that?
Think about it and you'll answer your own question there.
From a web developer perspective having statistics and trends on customer browser capabilities helps you design for the masses, ignore fading trends and on the flip side embrace upcoming trends.
As a matter of fact the Browser Capabilities Project [browsers.garykeith.com] provides a lot of information already that many websites use to target capabilities to the devices that connect to their sites.
A couple of of WebmasterWorld members already maintain a lot of these stats so the OP may want to work with them on this project.
My browser capabilities project is designed more to let developers know what browser-based features a browser is capable of supporting.
This commercial product provides a lot more information and lets developers know what actually is supported in both the browser and the underlying system.
[edited by: GaryK at 8:23 pm (utc) on July 6, 2009]
Actually, I am on a Linux server, so I'm afraid that I wouldn't be able to use that product
It works with Linux now with their SaaS version.
You're obviously using javascript to collect all of the information and now they're hosting the stats engine, you just plug in the javascript on your pages and that's all you need to do.
Thanks.